Skip to content
BranchesTags

Latest commit

 

History

5 Commits

Folders and files

NameName
Last commit message
Last commit date
backend
backend
 
 
data/processed
data/processed
 
 
frontend
frontend
 
 
ml
ml
 
 
models
models
 
 
.env.example
.env.example
 
 
.gitignore
.gitignore
 
 
README.md
README.md
 
 
openapi.json
openapi.json
 
 

Repository files navigation

KrishiMind

AI-Powered Crop Price Prediction and Agricultural Market Intelligence Platform

KrishiMind is an end-to-end AI-powered agricultural intelligence platform that helps users analyze crop market prices and forecast future mandi prices using LSTM-based deep learning models.

The platform combines time-series forecasting with agricultural market intelligence, including market price analysis, forecast confidence, mandi comparison, weather context, and crop-related news sentiment.

Live Demo

🌐 Frontend: https://krishimind-seven.vercel.app

🔗 Backend API: https://krishimind-api.onrender.com

📚 API Documentation: https://krishimind-api.onrender.com/docs

Problem Statement

Agricultural commodity prices can fluctuate significantly because of factors such as:

  • Market supply and arrivals
  • Seasonal demand
  • Weather conditions
  • Storage availability
  • Regional mandi variations
  • Agricultural news and market sentiment

Farmers and agricultural stakeholders often lack a single platform that combines historical market behavior with AI-based price forecasting and market context.

KrishiMind addresses this problem by providing crop-level price predictions and agricultural market intelligence through an interactive web platform.

Solution

KrishiMind uses trained Long Short-Term Memory (LSTM) models to forecast future crop mandi prices.

The prediction pipeline provides:

  1. Current market price
  2. Future predicted price
  3. Daily forecast values
  4. Forecast confidence
  5. Prediction error metrics
  6. Price trend analysis
  7. Advisory action
  8. Mandi comparison
  9. Weather context
  10. Agricultural news sentiment

The application transforms machine learning predictions into an interactive dashboard designed to make agricultural market forecasts easier to understand.

Key Features

AI Crop Price Prediction

  • LSTM-based time-series forecasting
  • Crop-specific trained models
  • Configurable prediction horizon
  • Multi-day forecast generation
  • Current vs predicted price comparison

Supported Crops

KrishiMind currently supports:

  • Onion
  • Tomato
  • Potato
  • Wheat
  • Rice

Forecast Analytics

Each prediction provides:

  • Current mandi price
  • Predicted future price
  • Percentage price change
  • Forecast trend
  • Daily forecast values
  • Upper prediction range
  • Lower prediction range
  • Model confidence
  • MAPE
  • MAE
  • RMSE

Agricultural Intelligence

The platform also provides contextual agricultural signals such as:

  • Market supply and arrivals
  • Weather conditions and impact
  • Agricultural news sentiment
  • Crop-specific market signals
  • Mandi comparisons

Mandi Comparison

Users can compare predicted crop prices across multiple mandis to understand regional market differences.

AI Advisory

Based on predicted market movement and available market context, KrishiMind generates action-oriented recommendations such as:

  • WAIT
  • HOLD
  • SELL

Prediction History

The dashboard stores generated predictions and displays information including:

  • Crop
  • Mandi
  • Prediction date
  • Forecast horizon
  • Predicted price
  • Current status
  • Accuracy
  • Recommended action

Interactive Dashboard

The frontend includes:

  • Crop selection
  • Prediction history
  • Price trend charts
  • Current vs predicted price comparison
  • Confidence visualization
  • Mandi comparison charts
  • Market signals
  • News sentiment
  • AI advisory

How KrishiMind Works

                    Historical Crop Price Data
                              │
                              ▼
                    Data Preprocessing
                              │
                              ▼
                    Feature Engineering
                              │
                              ▼
                   Crop-Specific LSTM Model
                              │
                              ▼
                      Price Forecast
                              │
             ┌────────────────┼────────────────┐
             ▼                ▼                ▼
       Daily Forecast    Error Metrics    Trend Analysis
             │                │                │
             └────────────────┼────────────────┘
                              ▼
                    Prediction API Response
                              │
               ┌──────────────┼──────────────┐
               ▼              ▼              ▼
        Weather Context   News Sentiment   Market Signals
               │              │              │
               └──────────────┼──────────────┘
                              ▼
                     KrishiMind Dashboard

System Architecture

┌─────────────────────────────────────┐
│          React Frontend             │
│                                     │
│  Dashboard │ Charts │ Predictions   │
│  Market Signals │ AI Advisory       │
└──────────────────┬──────────────────┘
                   │
                   │ REST API
                   ▼
┌─────────────────────────────────────┐
│          FastAPI Backend            │
│                                     │
│  Prediction Router                  │
│  Weather Service                    │
│  News Sentiment Service             │
│  Advisory Logic                     │
└──────────────────┬──────────────────┘
                   │
                   ▼
┌─────────────────────────────────────┐
│           ML Layer                  │
│                                     │
│  Crop-Specific LSTM Models          │
│                                     │
│  Onion                              │
│  Tomato                             │
│  Potato                             │
│  Rice                               │
│  Wheat                              │
└─────────────────────────────────────┘

Machine Learning Model

KrishiMind uses Long Short-Term Memory (LSTM) neural networks for crop price forecasting.

LSTM networks are suitable for time-series forecasting because historical price observations can influence future market prices.

Separate trained models are used for different crops.

The prediction pipeline provides:

  • Current price
  • Predicted price
  • Daily forecast
  • Forecast trend
  • Confidence range
  • MAPE
  • MAE
  • RMSE

Weather conditions and agricultural news sentiment are presented as contextual market intelligence signals alongside the model forecast.

Model Evaluation

The final KrishiMind evaluation was performed using 30 evaluation samples per crop.

Crop MAE RMSE MAPE
Onion 76.88 88.00 3.70%
Tomato 124.19 144.79 6.71%
Potato 312.80 345.42 20.92%
Rice 180.73 202.99 4.55%
Wheat 95.86 108.93 3.89%

Model performance varies across crops because each commodity has different levels of market volatility and price variation.

Technology Stack

Machine Learning

  • Python
  • TensorFlow
  • Keras
  • LSTM
  • Scikit-learn
  • Pandas
  • NumPy

Backend

  • Python
  • FastAPI
  • Pydantic
  • REST APIs
  • Uvicorn

Frontend

  • React
  • JavaScript
  • Vite
  • CSS
  • Data visualization charts

Deployment

  • GitHub
  • Render
  • Vercel

Project Structure

krishimind/
│
├── backend/
│   │
│   ├── routers/
│   │   ├── predict.py
│   │   ├── news.py
│   │   └── weather.py
│   │
│   ├── services/
│   │   ├── lstm_service.py
│   │   ├── news_service.py
│   │   └── weather_service.py
│   │
│   ├── main.py
│   └── requirements.txt
│
├── frontend/
│   │
│   ├── public/
│   │
│   ├── src/
│   │   ├── components/
│   │   │   ├── charts/
│   │   │   ├── prediction/
│   │   │   └── common/
│   │   │
│   │   ├── data/
│   │   ├── tabs/
│   │   └── utils/
│   │
│   ├── package.json
│   └── vite.config.js
│
├── data/
│
├── ml/
│
├── models/
│
├── .env.example
├── .gitignore
└── README.md

Local Setup

1. Clone the Repository

git clone https://github.com/snehapriy958/krishimind.git
cd krishimind

Backend Setup

2. Create a Virtual Environment

Windows

python -m venv venv
venv\Scripts\activate

3. Install Dependencies

pip install -r backend/requirements.txt

4. Configure Environment Variables

Create a .env file inside the backend directory.

Example:

NEWS_API_KEY=your_news_api_key
OPENWEATHER_API_KEY=your_openweather_api_key
GROQ_API_KEY=your_groq_api_key

5. Run the Backend

python -m uvicorn backend.main:app --host 127.0.0.1 --port 8002

For development with automatic reload:

python -m uvicorn backend.main:app --reload --host 127.0.0.1 --port 8002

The backend will be available at:

http://127.0.0.1:8002

API documentation:

http://127.0.0.1:8002/docs

Frontend Setup

1. Navigate to the Frontend Directory

cd frontend

2. Install Dependencies

npm install

3. Configure Environment Variables

Create a .env file:

VITE_API_URL=http://127.0.0.1:8002

4. Start the Development Server

npm run dev

The frontend will be available at:

http://localhost:5173

API Overview

KrishiMind exposes REST API endpoints for agricultural intelligence.

Endpoint Method Description
/ GET API status and information
/health GET Backend health check
/predict POST Generate crop price prediction
/weather/{state} GET Retrieve weather context
/news/{crop} GET Retrieve crop-related news sentiment

Interactive API documentation is available at:

https://krishimind-api.onrender.com/docs

Prediction Pipeline

User selects crop and market
            │
            ▼
Frontend sends prediction request
            │
            ▼
FastAPI validates request
            │
            ▼
Crop-specific LSTM model processes data
            │
            ▼
Future price forecast generated
            │
            ├── Daily forecast
            ├── Confidence range
            ├── MAPE
            └── Trend analysis
            │
            ▼
Weather context retrieved
            │
            ▼
News sentiment retrieved
            │
            ▼
Market advisory generated
            │
            ▼
Interactive dashboard visualization

Deployment

Frontend

The KrishiMind frontend is deployed on Vercel.

Live Application:
https://krishimind-seven.vercel.app

Backend

The FastAPI backend is deployed on Render.

Backend API:
https://krishimind-api.onrender.com

Health Check:
https://krishimind-api.onrender.com/health

API Documentation:
https://krishimind-api.onrender.com/docs

Current Capabilities

  • Multi-crop price forecasting
  • Five supported crops
  • LSTM-based predictions
  • Multi-day forecasts
  • FastAPI backend
  • React dashboard
  • Prediction history
  • Price trend visualization
  • Forecast confidence
  • MAPE, MAE and RMSE metrics
  • Mandi comparison
  • Weather context
  • Agricultural news sentiment
  • Market signals
  • AI-based advisory
  • Vercel frontend deployment
  • Render backend deployment
  • End-to-end frontend and backend integration

Future Improvements

Potential future improvements include:

  • Real-time mandi price ingestion
  • Automated model retraining pipelines
  • Additional supported crops
  • District-level price forecasting
  • Weather features directly integrated into model training
  • News sentiment directly integrated into model training
  • Persistent user accounts
  • Cloud-based prediction history
  • Price alerts and notifications
  • Mobile application support
  • Improved model explainability

Key Engineering Highlights

KrishiMind demonstrates a complete end-to-end AI application pipeline:

Historical Data
      │
      ▼
Data Preprocessing
      │
      ▼
Deep Learning Model
      │
      ▼
Model Evaluation
      │
      ▼
FastAPI Backend
      │
      ▼
React Frontend
      │
      ▼
Cloud Deployment

The project focuses not only on building machine learning models but also on integrating and deploying them as a usable full-stack AI application.

Author

Sneha Singh

AI/ML Engineer

License

This project is intended for educational, research, and portfolio purposes.

About

AI-powered agricultural intelligence platform for crop price forecasting and farmer decision support.

krishimind-seven.vercel.app

Resources

Readme
Activity

Stars

1 star

Watchers

0 watching

Forks

0 forks
Report repository

Releases

Packages

Contributors

Languages